Application of comorbidity indexes at baseline could be useful to predict rates of response in patients with chronic myeloid leukemia treated with imatinib

Abstract

In patients with cancer, comorbidities and poor performance status generally infl uence prognosis and aff ect therapeutic plans and outcome. In particular, comorbidities may have an impact on survival and choice of treatment among patients with cancer; comorbidities have been identifi ed as signifi cant determinants of response to therapy in older patients with acute myeloid leukemia, or female breast cancer, or head and neck or lung cancer [1]. Recently, we evaluated the impact of comorbidities in a large series of elderly patients with chronic myeloid leukemia (CML) treated with dasatinib after imatinib failure, with the aim of establishing associations between comorbidities and the development of pleural eff usions or compliance to the drug [2]. Th e prognostic importance of comorbidities in CML has never been evaluated in relation to active treatment. Th e aim of the present study was to compare and validate three diff erent non-specifi c comorbidity scores (Charlson Comorbidity Index [CCI] [3], Cumulative Illness Rating Scale [CIRS] [4] and Adult Comorbidity Evaluation 27 [ACE-27] [5]) by applying them to a cohort of patients with CML treated with imatinib as fi rst or second line after interferon (IFN) failure. All patients entered in this study were diagnosed and followed between 1995 and 2010 at the Sapienza University of Rome, outside of clinical trials. In all cases, diagnosis was confi rmed at the cytogenetics (conventional banding analysis, CBA) and the molecular level (reverse trancriptase [RT]- and real-time quantitative polymerase chain reaction [RQ-PCR] for identifi cation of Philadelphia chromosome positivity (Ph ) and BCR/ABL1 transcript). Of 343 patients enrolled in this study, 142 fi rst received interferon- α outside of clinical trials and were then switched to imatinib for failure; for this group of patients, data relating to comorbidities were collected at the time of the start of imatinib. Th e remaining 201 patients were treated consecutively with imatinib as fi rst line from January 2000 onward. Risk evaluation at baseline was performed using the Sokal score, and retrospectively with the European Treatment and Outcome Study (EUTOS) score. All patients were followed according to European LeukemiaNet (ELN) guidelines from 2006 to date [6]. Cytogenetics analysis was performed after 3 months from the start and then at 6 and 12 months, and then every year after a complete cytogenetic response (CCyR) was achieved. Results were expressed according to ELN defi nitions, according to CBA fi ndings. Molecular testing was performed with RQ-PCR every 3 months using peripheral blood, and then every 6 months after a major molecular response (MMR) or a complete molecular response (CMR) was achieved. Results were expressed as a ratio relative to an ABL1 control, with the fi nal value converted to the International Scale (IS). We considered as primary resistance the lack of any cytogenetic response, whereas secondary resistance was the achievement and subsequent loss of cytogenetic or molecular response. Application of the CCI revealed that 182 patients had score 0 – 2 (50%), 101 patients had score 3 (29%), 47 patients had score 4 (13.7%) and 13 patients had score 5 (3.7%). We found a signifi cant correlation between stratifi cation according to the CCI and cumulative incidence of CCyR ( p 0.02, Figure 1), cumulative incidence of MMR ( p 0.03) and cumulative incidence of primary and secondary resistance ( p 0.01, Table I). Application of the ACE-27 index in 338 evaluable patients identifi ed 170 patients (50%) as having score 2 and 168 patients (49.7%) as having score 3; we did not reveal any substantial diff erence according to this stratifi cation in terms of CCyR, MMR or resistance (Table I). Finally, we applied the CIRS score in 334 patients, which identifi ed 155 patients as having score 1 (46%), 96 patients as having score 2 (29%), 61 patients as having score 3 (18%), 20 patients as having score 4 (6%) and two patients as having score 6 (0.6%). Again, as for the ACE-27 index, we did not fi nd any signifi cant association between CIRS stratifi cation and cumulative incidences of CCyR, MMR and resistance. We compared imatinib dose intensity, temporary dose interruption and permanent discontinuation according to the CCI. We found a signifi cant correlation between higher scores of the CCI and low compliance to treatment due to the higher rate of toxicity and consequently higher rate of temporary (21% in patients with score 0 – 2 vs. 56% in patients with score 3, p 0.01) and permanent discontinuation (5%